In patches of rural Appalachia and the Rust Belt, health authorities were sounding alarms that a powerful painkiller called Opana had become the drug of choice among people abusing prescription pills.

It was twice as potent as OxyContin, the painkiller widely blamed for sparking the opioid crisis, and was relatively easy to dissolve and inject. By 2015, government investigations and scientific publications had linked its misuse to clusters of disease, including a rare and life-threatening blood disorder and an H.I.V. outbreak in Indiana.

Opana’s manufacturer, the pharmaceutical company Endo, had scaled back promotion of the drug. But months later, the company abruptly changed course, refocusing resources on the drug by assigning more sales representatives.

The push was known internally as the “Sales Force Blitz” — and it was conducted with consultants at McKinsey & Company, who had been hired by Endo to provide marketing advice about its chronic-pain medicines and other products.

The untold story of McKinsey’s work for Endo was among the revelations found by The New York Times in a repository of more than 100,000 documents obtained by a coalition of state attorneys general in a legal settlement related to McKinsey’s opioid work.

Much has been disclosed over the years about McKinsey’s relationship with Purdue Pharma, including the consulting firm’s recommendation that the drug maker “turbocharge” its sales of OxyContin. But The Times found that the firm played a far deeper and broader role in advising clients involved in the opioid crisis than was publicly disclosed.

The McKinsey records include more than 15 years of emails, slide presentations, spreadsheets, proposals and other documents. They provide a sweeping and detailed depiction of a firm that became a trusted adviser to companies at the core of an epidemic that has claimed half a million American lives.

While the firm held remarkable sway at Purdue, it also advised the largest manufacturer of generic opioids, Mallinckrodt. It worked with Endo on marketing Opana and helped it grow into a leading generics manufacturer. It advised Johnson & Johnson, whose subsidiary Tasmanian Alkaloids was the largest supplier of the raw materials extracted from poppies used to make many top-selling opioids. Then, as the full brunt of the epidemic became apparent, it counseled government agencies on how to address the fallout.

McKinsey’s opioid clients already wanted to grow their businesses. What the firm offered was know-how and sophistication, the documents show, and, as it noted in one presentation, “in-depth experience in narcotics.”

The Massachusetts attorney general, Maura Healey, who helped craft the settlement, said in a statement that “as Americans were dying from the opioid epidemic, McKinsey was trading on its reputation and connections to make the crisis worse.” She added that the newly released documents “expose McKinsey’s role in the opioid crisis and will inform policymakers’ efforts to prevent this from happening again.”

Drawing on reams of data and proprietary tools, McKinsey vetted deals and advised on corporate strategy. It developed tactics for dealing with regulators and helped secure approval for new products.

The firm helped clients adopt more aggressive sales strategies, which, on at least two occasions, led companies to shift resources to more potent products. It profiled and targeted physicians, in some instances trying to influence prescribing habits in ways that federal officials later warned heightened the risk of overdose.

And when opioid prescriptions began to decrease during a government crackdown, the records show, McKinsey devised new approaches to drive sales.

McKinsey agreed to provide the documents to the attorneys general last year as part of a nearly $600 million settlement in which it admitted no wrongdoing. The firm has since apologized for its advice to opioid makers but, in a statement on Wednesday, suggested that its work with companies other than Purdue was “much more limited” and that it “did not counsel or recommend to Endo that it promote Opana more aggressively.”

“We recognize the terrible consequences of the opioid epidemic and have acknowledged our role in serving opioid manufacturers,” said a McKinsey spokesman. “We stopped that work in 2019, have apologized for it and have been focused on being part of the solution.”

An Endo spokeswoman declined to comment on the company’s work with McKinsey, citing litigation. She instead referred to a company statement saying that in September 2016 Endo had “stopped promoting opioid products to health care professionals” and eliminated its opioid-focused sales force.

Mallinckrodt declined to comment. Johnson & Johnson, in a statement, maintained that all its actions were appropriate, while Purdue said that it was focused on ending bankruptcy proceedings so it could reorganize into a new, more “public-minded” company that would “deliver billions of dollars of value” toward abating the opioid crisis and compensating victims.

‘Opana Patients’

Dr. Steven Butler, a kidney specialist serving a largely rural stretch of East Tennessee, helped with an unusual case in fall 2012. A woman in her 20s had arrived at the Holston Valley Medical Center in Kingsport with an array of symptoms — she was anemic, and her kidneys appeared to be failing — that resembled a rare blood disorder.

A few days later, another patient with similar symptoms arrived at the hospital. Then a third. Dr. Butler called the Tennessee Department of Health, which launched an investigation. Over the following months, more patients appeared.

As they underwent time-consuming treatments, some acknowledged they had dissolved and injected a pill whose name Dr. Butler had never heard before: Opana ER.

“Locally, it became a very well-described phenomenon,” he recalled. “They were just called ‘Opana patients,’ as though that was a real common thing.”

The tangled path that led to Opana’s rise illustrates McKinsey’s deep involvement in the opioid business, with its work for one client rippling out with consequences for others.

Years earlier, the firm had helped usher the drug onto the market, advising Endo’s partner, Penwest Pharmaceuticals, on its launch in 2006. Two years later, the documents show, McKinsey performed a project for Purdue that paved the way for Endo to extend Opana’s reach.

Purdue was seeking approval from the Food and Drug Administration for a new version of OxyContin that would be more difficult to snort or inject. After the F.D.A. denied its application in 2008, Purdue enlisted McKinsey’s help. The consultants interviewed a former drug dealer about OxyContin abuse, oversaw scientific studies, prepared regulatory documents and coached company officials on how to deal with the F.D.A., which had been a McKinsey client. The agency gave its approval in 2010, and later allowed Purdue to claim the new pills were resistant to abuse.

Soon, OxyContin sales declined — while Opana sales rose. In an internal document, Endo attributed the uptick in part to “patient dissatisfaction with new OxyContin formulation.” Data on abuse showed similar trends: a decline for OxyContin and a rise for Opana.

Endo later developed a new version of Opana it wanted to promote as abuse-resistant. The F.D.A. found that the new pills “demonstrated a minimal improvement in resistance to tampering by crushing,” and that they were “readily abusable” by injection. The agency allowed the drug to enter the market in early 2012, but without being labeled as resistant to abuse.

Within months, Dr. Butler saw his first Opana patient. In October 2012, both the F.D.A. and the Centers for Disease Control and Prevention put out health alerts about the blood syndrome. Then another cluster appeared, in North Carolina, and other cases in Arkansas, Florida, Pennsylvania and South Carolina.

To make matters worse, according to the F.D.A., the new version of Opana drove many users to switch from snorting to injecting, considered a riskier form of abuse. The likely cause of the blood disorder, researchers determined, was the very substance that Endo had added to make the pills harder to crush. When dissolved and injected, it could trigger rapid red blood cell destruction and organ damage.

As concerns about Opana grew, Endo hired a new chief executive in 2013: Rajiv De Silva, a former leader within McKinsey’s pharmaceutical practice who soon tapped the firm to help chart a growth strategy.

A few months after Mr. De Silva took over, McKinsey helped Endo execute a complicated maneuver known as a “tax inversion” — a legal form of tax avoidance that the Obama administration would decry as an “abuse” of the system. For tax purposes, the Pennsylvania company was now based in Ireland, where the rate was substantially lower.

The move, which sent Endo’s stock price climbing, was “a tax play to set up doing a lot of deals,” according to a 2014 email from a McKinsey partner named Dr. Arnab Ghatak, who also helped lead the firm’s work with Purdue.

Endo went on a buying spree and would soon become one of the largest U.S. manufacturers of generic opioids.

‘The Narcotics Franchise’

The production of pills by companies like Endo and Purdue depended on a complex and tightly regulated global supply chain stretching from the fields of Tasmania to factories in the American heartland.

Here, too, was McKinsey.

Long before a patient in the United States filled a prescription for OxyContin, a farmer on another continent harvested a poppy rich in a substance called thebaine. Tasmanian Alkaloids, the Johnson & Johnson subsidiary, controlled the majority of this market.

From far-flung fields and extraction facilities, the raw materials made their way to American processing plants. The top U.S. producers at this stage were another Johnson & Johnson subsidiary, Noramco, and Mallinckrodt, the big generics manufacturer.

The documents reveal McKinsey’s work advising them behind the scenes. By the firm’s own account, it had deep expertise in the international trade of legal narcotics. “We serve the majority of the leading players,” the consultants wrote in a 2009 memo.

That year, the firm oversaw a project for Johnson & Johnson titled “Maximizing the Value of the Narcotics Franchise.” In a presentation set against an image of a poppy field, the consultants advised the company on how it could invest to further strengthen its already-dominant position or sell the business if the price was right.

For Mallinckrodt, McKinsey consultants walked factory floors and monitored production data, recommending how the company might coax greater yields from the same base of raw materials and speed up manufacturing lines.

In 2016, McKinsey prepared Mallinckrodt for negotiations with companies that sourced generic drugs for Walmart and CVS, and advised on dealing with the Drug Enforcement Administration. The D.E.A. had set production limits to prevent an oversupply of pills, and McKinsey counseled Mallinckrodt on how it could use logistical tactics to secure a higher quota while maintaining a “friendly relationship” with the agency.

“To suggest this work was intended to undermine relevant laws or regulations would be false,” the McKinsey spokesman said.

McKinsey consultants also took jobs at the opioid manufacturers themselves. A partner in the firm’s pharmaceutical practice, Frank Scholz, became Mallinckrodt’s senior vice president of global operations in 2014 and later was promoted to president of its generics business.

But it was the arrival of Mr. De Silva at Endo that brought a particular opportunity for McKinsey. In late 2014, the company asked the consultants to provide advice on structuring the company’s sales force. This soon evolved into a more detailed project in an area where McKinsey excelled: how to dispatch hundreds of sales representatives to maximum effect.

Shifting to Offense

McKinsey had a playbook for seemingly any problem a pharmaceutical company might face, from production snags to generic competition to inquisitive regulators. But the firm had a particular penchant for sales and marketing.

In the years leading up to its work on Opana, McKinsey had built increasingly powerful tools for getting the right messages in front of the right physicians, and the firm had honed them in numerous opioid-marketing projects, including two for Johnson & Johnson.

While the broad strokes of these efforts have been known, the documents provide an unprecedented look inside McKinsey’s tool kit. The records related to the firm’s work for Purdue are particularly detailed, providing insight into the strategies that consultants used for other companies.

In 2009, the firm recommended a technique known as segmentation. The best marketing campaigns — whether for food, cars or electronics — divided consumers into segments based on how they acted and thought, then developed tailored messages to win them over, the consultants said.

In Purdue’s case, the customer was a physician with a license to prescribe controlled substances, and the product was OxyContin.

The consultants interviewed dozens of physicians and solicited the views of hundreds more in a survey. Four groups of doctors emerged, each with a distinct profile. The consultants then developed messages to appeal to each group’s practical and emotional needs.

McKinsey identified a particular opportunity in doctors who were hesitant to prescribe OxyContin because of worries about abuse, addiction and possible scrutiny from the D.E.A. These physicians often tried to treat chronic pain with less powerful drugs.

Persuading them to switch to OxyContin could be worth hundreds of millions of dollars, McKinsey advised. To do this, McKinsey proposed tactics to “raise physician comfort levels through appropriate education and support.” Sales representatives, McKinsey said, should reassure doctors that many of their colleagues prescribed OxyContin and that the drug need not be reserved for extreme pain.

In 2014, the F.D.A. introduced new labeling requirements for OxyContin and similar opioids, limiting their use to cases of severe chronic pain in which less risky treatments had proved ineffective. But McKinsey’s strategy had long since been rolled out.

Another McKinsey approach, known as targeting, tried to identify doctors who would provide the greatest return on sales representatives’ time.

Purdue, dissatisfied with dipping OxyContin sales in 2013, had enlisted McKinsey’s help. Revenues were down, the consultants advised, in large part because of government actions to tamp down the opioid epidemic. Doctors were writing prescriptions for fewer tablets and lower doses, and wholesalers and pharmacies were imposing new controls.

McKinsey recommended a more aggressive response than the one Purdue’s vice president for sales and marketing, Russell Gasdia, had been pursuing. Mr. Gasdia had accepted that OxyContin revenue was dropping in part “due to less abuse,” one McKinsey consultant wrote, and he was focused on promoting a less potent opioid.

McKinsey called for a shift “to offense”: Purdue needed physicians to start new patients on OxyContin. Drawing on an array of data — more than just a list of high prescribers, which had been the focus of Purdue and other drug companies — the consultants identified specific doctors to target.

In a statement, McKinsey said that this advice pertained to the reformulated OxyContin, which “was believed to be a safer version of the product.”

Purdue’s board endorsed the plan, and soon Mr. Gasdia stepped down as head of sales and marketing. In an internal self-assessment, Dr. Ghatak, the McKinsey partner who helped lead its Purdue business, basked in the firm’s success.

“Overall,” he wrote, “we are now deeply involved in nearly every facet of the company.”

An Outbreak and a Blitz

When a cluster of H.I.V. cases appeared in a small southeastern Indiana community in 2015, it didn’t take the C.D.C. long to identify the cause. Most of the patients had injected Opana.

The governor declared a public health emergency, and the list of those infected eventually surpassed 180.

Disease often followed incidents of injecting opioids, but Opana posed a higher risk, the C.D.C. later determined. When injected, it was 10 times as potent as morphine. The high was intense but short-lived, and the withdrawal was particularly agonizing. As a result, users injected more frequently.

And because Opana commanded a high street price, users often split pills, shared equipment and shot up multiple times in one sitting. It was a recipe for what a C.D.C.-led research team called “explosive transmission.”

If any of this caused alarm among the McKinsey consultants working for Endo, their presentations did not reflect it.

In summer 2015, McKinsey helped launch the “Sales Force Blitz,” which the firm said in a statement applied to a range of Endo’s products. “The small portion of our work that concerned Opana was done at the client’s request,” the spokesman said, “not by our recommendation.”

While the company had pulled back its marketing of the painkiller, McKinsey now advised it on how to do the opposite, emails and presentations show, by reallocating sale representatives from a migraine drug to Opana.

A consultant, Sherin Ijaz, expressed her excitement in an email to the head of Endo’s pain business unit, John Harlow. The next step “is to identify the sweet spot of docs so we can do targeting,” she wrote, adding that the “fun” begins “on Monday!”

“Agreed,” Mr. Harlow replied, “and the fun is just beginning!”

When two Endo executives proposed shifting some sales calls to promote the company’s arthritis gel, McKinsey was opposed. Doing so would be a distraction “at a time when we want to drive Opana,” wrote another McKinsey consultant, Nicholas Mills.

Ultimately, the consultants directed Endo to focus on more than 3,000 additional physicians with promotional messages about Opana.

In 2017, less than two years later, the F.D.A. took the rare step of demanding that Endo pull Opana from the market, citing the grave public health consequences of its abuse. The company complied.

Over the five years from the appearance of the blood-disease cluster in Tennessee to the drug’s withdrawal from the market, the painkiller had brought in more than $844 million in revenue, according to corporate filings.

In Indiana, law enforcement officials broke up a drug-trafficking ring in 2016. One man admitted obtaining Opana in Detroit and selling it in bulk to a dealer. He was sentenced to six years in prison.

“Health care, the schools, the welfare department, the whole thing is crumbling because of drugs, drugs that you helped make available,” said the judge in the case, scolding him.

“You’re not responsible for all of that, of course, but you did your part.”

‘Opioid Crisis Is Horrible’

In June 2017, Tom Latkovic rose to speak at a health care conference in Chicago sponsored by his employer, McKinsey.

“I start today by asking, ‘Why do we continue to prescribe, dispense, pay for opioid prescriptions to people that we know, or at least we could know, have an incredibly high propensity to abuse them?’”

Mr. Latkovic, a senior partner, was not a member of McKinsey’s pharmaceutical practice. Instead, his team focused on using data analysis tools to address complex health care problems, and it had increasingly homed in on the opioid epidemic.

In the hopes of broadening this work, Mr. Latkovic told the audience, “We are launching a new center focused on opioids and insights.”

The client list for the new venture came to include state governments, insurers and health systems. One of McKinsey’s more ambitious efforts was in Philadelphia, a city that had one of the highest death rates in the country from opioid overdoses.

In 2019, consultants spent almost two months working with the city government, according to two people who were local officials at the time. Both praised McKinsey’s work, which came at no cost to the city but was later shelved after Covid-19.

Yet as Mr. Latkovic’s team tried to combat the opioid epidemic, the firm did not stop serving the company often blamed for sparking it, Purdue. And on at least two occasions, the documents show, drafts of publications prepared by Mr. Latkovic’s team were given to consultants for pharmaceutical clients to review. The purpose, a manager in the pharmaceutical practice wrote, was to assess “whether this could create any waves on social media or from journalists that could be harmful to our Pharma clients.”

As negative news coverage and lawsuits against Purdue mounted, some of the consultants fretted internally that scrutiny might extend to McKinsey.

In 2019, around the time of the Philadelphia project, McKinsey decided to stop advising companies on opioids — after the firm’s 15-year relationship with Purdue became public as part of a court filing by the Massachusetts attorney general’s office. Since Mr. Latkovic’s 2017 speech, McKinsey had collected $7.8 million in fees from Purdue, the documents show.

The disclosure that McKinsey had advised Purdue led to debate within the firm. “We may not have done anything wrong, but did we ask ourselves what the negative consequences of the work we were doing was, and how it could be minimized?” one consultant wrote.

Dr. Ghatak, a driving force behind McKinsey’s work for Purdue and Endo, found himself in the spotlight. Much as he had done for pharmaceutical executives, he crafted talking points, this time for himself.

“Opioid crisis is horrible,” he wrote. “Acknowledge that up front.” But by advising clients to develop products that would be more difficult to abuse, “we were directly working on a solution to a public health crisis, not a silver bullet but definitely a solution.”

In 2020, documents released as part of a Purdue legal case indicated that Dr. Ghatak and another consultant, Martin Elling, had discussed destroying records. McKinsey soon fired them.

The firm settled with the state attorneys general in early 2021.

Some of McKinsey’s former clients faced potentially crushing damages in court. Purdue filed for bankruptcy protection in 2019, and Mallinckrodt did the same the following year. Johnson & Johnson had previously sold its narcotics business to a private investment firm and has settled a number of lawsuits related to its marketing of opioids, which the company said in a statement was “appropriate and responsible.”

Endo has also floated the possibility of bankruptcy amid a wave of litigation over its marketing of opioids, especially Opana. The company said in a regulatory filing that it had received a subpoena in 2020 from the U.S. attorney’s office for the Western District of Virginia, which years earlier had won guilty pleas from Purdue executives. This time, according to Endo’s disclosure, the office wanted information on McKinsey.

Top illustration by Mark Weaver.